Frequency tuning mechanisms are required in a wide range of different applications. For example they are required in wireless transceivers for the down conversion of signals at different frequencies, for multiple-band applications and for wide band applications. An ideal frequency tuning circuit will have a wide tuning range, be power efficient and have a high operating frequency.
Conventionally a tuning circuit includes an LC tank, and a conventional method of providing control of the tuning frequency is by using a variable capacitance such as a varactor to vary the value of C in the LC tank. Several classes of varactors, such as junction diodes and MOS capacitors, are commonly found. However, this arrangement has the disadvantage that there is a limited frequency range (about 10% only) owing to the limited capacitance ratio of the varactor. Also, because the power consumption of the LC tank is high, it is more power efficient to minimize the capacitance of the tank and to adjust the inductance for frequency variation. As a result, it is highly desirable to be able to implement integrated variable inductors.
Currently, several techniques are available for providing a variable inductance. These include active inductors and switched resonators. A typical design for an active inductor is the gyrator-C architecture, which employs a gyrator and an integrating capacitor. A gyrator consists of two transconductors connected in a feedback configuration, as shown in FIG. 1. This type of active inductor makes use of the parasitic capacitance of the transistors as the integrating capacitor. The inductance of active inductor is:
            Z              i        ⁢                                  ⁢        n              ⁡          (              j        ⁢                                  ⁢        ω            )        =                    g        ds1            +              jω        ⁡                  (                                    c              gs2                        +                          c              gd1                        +                          c              gd2                                )                                    (                              g            m1                    +                      g            ds1                    +                      j            ⁢                                                  ⁢            ω            ⁢                                                  ⁢                          c              gd2                                      )            ⁢              (                              g            m2                    +                      jω            ⁡                          (                                                c                  gs2                                +                                  c                  gd1                                            )                                      )            
Because only a few active devices are used in this type of inductor, the chip area occupied is usually very small. Tunability is another advantage of this type of active inductor. As shown in the above equation, by changing the bias and, therefore, the transconductance of the transistors, the inductance of the active inductor can be varied.
However, the power consumption and noise contribution of the active devices used in these inductors are generally too high to be practical, and the dynamic range is quite limited. Most important of all, active inductors are generally not suitable for high frequency operation. At high frequencies, the performance of the active inductor is degraded by the phase errors induced by parasitics.
Recently, switched resonators using multiple inductors have been introduced. A switched resonator typically comprises two spiral inductors and a switching transistor, either connected in parallel or in series with the inductors as shown in FIG. 2. If the switching transistor is connected in parallel with one of the inductors, the inductor is shorted when the switch is on. As a result, the equivalent inductance reduces from L1+L2 to L1.
A switched resonator can be used for coarse tuning and another varactor can be used for fine tuning. The tuning range of the resonator can therefore be significantly improved. However, the turn-on resistance of the switching transistor has a great impact on the quality factor of the resonator. It is necessary to increase the size of the transistor in order to reduce the effect of the turn-on resistance on the quality factor. Since the operating frequency of the resonator depends on the equivalent inductance and the capacitance between drain and ground of the switching transistor, the drain capacitance of the switch significantly reduces the operating frequency of the resonator. Thus, this type of switched resonator is not suitable for applications with low noise, low power, and high frequency.
It is also possible to frequency tune some types of resonators by mechanically changing a property of these resonators. However, this is not feasible if the resonator is to be integrated on chip.
Thus, it is desirable to provide a variable inductance that is suitable for high frequency circuits and has reduced power consumption and reduced noise degradation.